Devesa J, Avedillo D, Morales V, Nunez Puertas A
Surg Gynecol Obstet. 1984 May;158(5):482-7.
The Dukes' and TNM systems for staging carcinoma of the colon and rectum are still the best pathologic classifications, but they do not apply to all patients and do not distinguish between patients who will die and patients who will be cured by the same therapeutic procedure. A new approach to this problem should be to establish a biochemical automatic classification, complementary to the morphologic one, which allow us to classify every patient before and after the first and subsequent treatments. By using several nonspecific tumor markers, such as CEA, AAT, AF, AAG, GGT and transferrine, a discriminant analysis was executed among the groups of patients with LD, RD and DD. Our initial results with only 12.8 per cent of incorrect classifications, that is patients classified in a less advanced group, suggest that this system may be quite useful in order to select those patients with carcinoma of the rectum who should benefit from preoperative radiotherapy as well as those who should receive adjuvant therapy after the first treatment. On the other hand, for patients classified in a more advanced group than the pathologic grading, we may well be able to identify those patients with occult disease for which the frequency of revisions should be shorter.
用于结肠癌和直肠癌分期的Dukes' 分期系统和TNM分期系统仍然是最佳的病理分类方法，但它们并不适用于所有患者，也无法区分接受相同治疗方法后会死亡的患者和会治愈的患者。解决这个问题的新方法应该是建立一种与形态学分类互补的生化自动分类方法，使我们能够在首次及后续治疗前后对每个患者进行分类。通过使用几种非特异性肿瘤标志物，如癌胚抗原（CEA）、α1抗胰蛋白酶（AAT）、甲胎蛋白（AF）、α1抗糜蛋白酶（AAG）、γ-谷氨酰转移酶（GGT）和转铁蛋白，对局限性疾病（LD）、区域性疾病（RD）和远处转移疾病（DD）患者组进行了判别分析。我们最初的结果显示，错误分类率仅为12.8%，即将患者分类到较早期组，这表明该系统在选择那些应从术前放疗中获益的直肠癌患者以及那些在首次治疗后应接受辅助治疗的患者方面可能非常有用。另一方面，对于分类结果比病理分级更晚期的患者，我们很可能能够识别出那些隐匿性疾病患者，对于这些患者，复查频率应该更短。
